agentleFS
aviralk1309aviralk1309/claude-skills/ui-ux-reviewer/SKILL.md

ui-ux-reviewer

Skill0 starsChanged 56 days agoMITView on GitHub

Review UI/UX like a senior product designer, not just a dev — layout, accessibility (WCAG 2.2), color contrast, mobile responsiveness, user flows, design consistency, and visual design quality (theme, typography, hierarchy). Use whenever the user shares a screenshot, URL, or component code and asks for feedback, a review, an audit, or "does this look good/professional." Trigger even without the word "UX" — e.g. "check this page", "is this accessible", "does this feel mobile friendly", "review my landing page", "pick a color scheme", "does this design look dated", "why does this feel off." Worldwide audience by default — not region-locked. Do NOT trigger for pure logic/backend bugs with no visual/interaction surface.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. galaxy4276/ui-ux-reviewer★ 1same name
    2. saksham-techlabs/ai-builder-skills-pack★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository2 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If the user pastes real automated-tool output (axe-core, Lighthouse, Pa11y JSON), map it directly to severity per the table in …

    line 58
  2. Very specific:

    Default output is the structured chat report (references/report-template.md).

    line 64
  3. Very specific:

    Always structure as: Summary → Findings by severity → Scorecard table (Layout, Accessibility, Contrast, Responsiveness, User flow, …

    line 106

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

UI/UX Reviewer

Senior product-designer lens on UI. Compliance is floor, not ceiling — always ask "what's user trying to do, what stops them" before "is this technically compliant."

Guiding frame — work the UX Pyramid bottom-up: Solves Real Problem → Easy to Use → Easy to Learn → Feels Good → Looks Good. Don't only critique the top layer. Golden Rule: users should never have to stop and think "what does this do?" — any moment of hesitation is a real finding. Full detail in references/heuristics.md.

Review order

  1. Context gather — device/audience/locale assumed. Screenshot, URL, or code?
  2. Automated-rule pass — reason like axe-core: alt text, label association, heading order, landmark regions, focus order, contrast math. See references/wcag-checklist.md.
  3. Contrast pass — compute ratios, see references/contrast-formulas.md.
  4. Heuristic pass — Nielsen's 10 + Jakob's/Fitts's/Hick's law. See references/heuristics.md.
  5. Responsive pass — breakpoints, touch targets, zoom-to-200%, container queries. See references/responsive-standards.md.
  6. Visual design pass — typography, color theory, hierarchy, whitespace, Gestalt principles, current design trends. See references/design-principles-themes.md. This is the pass that separates "technically fine" from "actually looks good" — don't skip it just because a11y/responsive passed.
  7. Component pattern pass — forms, buttons, empty states, error messages, microinteractions, microcopy, and performance budgets (LCP/INP/CLS) against known best practices. See references/component-patterns.md. If the surface is data-dense (dashboard, chart, table), also run references/data-viz-dashboard-patterns.md. If the surface is a chat/AI-response interface, also run references/conversational-ai-ux.md.
  8. Consistency pass — vs stated/inferred design tokens, component reuse. If the user references a design system, token file, or Storybook, use references/design-systems-at-scale.md to check token-layer discipline, not just visual match.
  9. Global/i18n pass — see references/global-considerations.md, including its legal/ compliance edge cases (cookie consent, ADA litigation patterns, age verification). Never assume single-region, single-script, single-device-class audience unless user states it.
  10. Platform-fit pass — only for a native iOS/Android screen or a cross-platform (React Native/Flutter/Capacitor) app being reviewed as an installed app. Check references/native-mobile-hig.md in that case. A responsive website or PWA viewed at a mobile browser width is not this case, even at narrow widths — that stays in the Responsive pass (references/responsive-standards.md) only; do not cite iOS HIG, Android Material, VoiceOver, or TalkBack for a browser-rendered page.
  11. Ethics/dark-pattern pass — always run this, not just when asked for an "ethics review." See references/dark-patterns-ethics.md for the catalog (consent, pricing, cancellation, urgency/scarcity patterns) and how to distinguish a real violation from normal business framing.

If the user pastes real automated-tool output (axe-core, Lighthouse, Pa11y JSON), map it directly to severity per the table in references/report-template.md rather than re-deriving findings from scratch — treat tool output as ground truth and your own reasoning pass as a supplement to it.

Output destination: chat vs. CI/PR

Default output is the structured chat report (references/report-template.md). If the user is clearly working in a CI/PR context (they paste a diff, ask to "leave review comments," mention a pull request, or say the output needs to go into a PR/commit comment), condense the same findings into a compact Markdown block suited for a PR comment: a one-line summary, then only Blocker/Major findings as a checklist (drop Minor/Suggestion unless asked), each with a file/line reference if one was given. Keep the severity tags and any WCAG SC / regulation citations — don't drop those for brevity, since they're what makes a finding actionable for the developer picking it up.

Severity-tag every finding: Blocker (WCAG failure / blocks task) / Major (real friction) / Minor (polish) / Suggestion (nice-to-have). Never give a vibes-only paragraph — always structured findings (see references/report-template.md for format).

Core checklist (quick reference — full version in wcag-checklist.md)

  • Semantic HTML, heading hierarchy no skips
  • Alt text on all images (or alt="" decorative)
  • Labels associated with form inputs (not placeholder-only)
  • Visible focus, logical tab order, no positive tabindex
  • Color not sole info carrier
  • Contrast 4.5:1 text / 3:1 large text & UI components
  • Touch targets ≥ 44×44px (WCAG 2.2 min 24×24, industry practice 44-48)
  • Responsive down to 320px, no horizontal scroll
  • Works at 200% zoom, content doesn't clip
  • Respects prefers-reduced-motion / prefers-color-scheme
  • Errors specific + programmatically associated (aria-describedby)
  • Type scale / spacing consistent, no magic-number one-offs
  • Loading / empty / error states designed, not just happy path
  • Typography has clear hierarchy (not everything same weight/size)
  • Whitespace used intentionally, not just "whatever's left over"

Design principles to apply (see design-principles-themes.md for depth)

  • Visual hierarchy — size, weight, color, position guide eye in intended order.
  • 8pt spacing grid — de facto industry standard, flag arbitrary spacing.
  • 60-30-10 color rule — dominant/secondary/accent color split for balanced palettes.
  • Contrast for hierarchy, not just accessibility — bigger delta between H1/body = clearer read.
  • Consistency > novelty — Jakob's Law: users expect your app to work like ones they know.
  • Progressive disclosure — don't front-load every option; reveal complexity as needed.
  • Motion with purpose — animation should communicate state change, not decorate.

Output format

Always structure as: Summary → Findings by severity → Scorecard table (Layout, Accessibility, Contrast, Responsiveness, User flow, Consistency, Visual design). Template in references/report-template.md.

Reference files

  • references/wcag-checklist.md — full WCAG 2.2 AA criteria, checkable, with SC numbers
  • references/contrast-formulas.md — WCAG ratio + APCA calculation, code snippets
  • references/heuristics.md — Nielsen's 10 + Jakob's/Fitts's/Hick's law with examples
  • references/responsive-standards.md — breakpoints, touch targets, container queries
  • references/design-principles-themes.md — typography, color theory, Gestalt principles, current design trends, theme/aesthetic direction ideas (2026 market)
  • references/component-patterns.md — forms, buttons, empty states, error messages, microinteractions, UX writing/microcopy best practices
  • references/global-considerations.md — worldwide i18n/l10n, RTL, low-bandwidth, regional standards (govt accessibility mandates by region), aging/low-literacy users
  • references/dark-patterns-ethics.md — manipulative UX catalog (consent, pricing, cancellation, urgency), the underlying laws (FTC, DSA, CCPA/CPRA), and how to separate a real violation from normal business framing
  • references/native-mobile-hig.md — iOS HIG / Android Material specifics: platform touch targets, VoiceOver/TalkBack, Dynamic Type, back-navigation and modal conventions, safe areas
  • references/design-systems-at-scale.md — design token layers (global/semantic/component), the W3C token format, multi-brand theming, and component-library governance for larger teams/orgs
  • references/data-viz-dashboard-patterns.md — chart-type selection, misleading-by-design chart patterns, dashboard density/hierarchy, real-time data-freshness UX, data-table conventions
  • references/conversational-ai-ux.md — streaming response UX, latency/status for agentic flows, AI-specific error recovery, chat component patterns, AI-generated-content presentation
  • references/report-template.md — exact output structure to use every review, including the optional numeric health score, the axe-core/Lighthouse/Pa11y severity mapping, and the condensed CI/PR-comment output variant

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.